Complementing Dynamical Downscaling With Super‐Resolution Convolutional Neural Networks

Deeksha Rastogi, Haoran Niu, Linsey Sara Passarella, Salil Mahajan, Shih‐Chieh Kao, Pouya Vahmani, Andrew D. Jones

Geophysical Research Letters · 12 February 2025 · Lawrence Berkeley National Laboratory / NERSC, Oak Ridge National Laboratory, University of California, Berkeley

Paper PDF

In one sentence. SRCNN trained on dynamically downscaled precipitation emulates high-resolution downscaling, with quantile and exponential losses improving extremes and GCM applicability.

Abstract

Abstract Despite advancements in Artificial Intelligence (AI) methods for climate downscaling, significant challenges remain for their practicality in climate research. Current AI‐methods exhibit notable limitations, such as limited application in downscaling Global Climate Models (GCMs), and accurately representing extremes. To address these challenges, we implement an AI‐based methodology using super‐resolution convolutional neural networks (SRCNN), trained and evaluated on 40 years of daily precipitation data from a reanalysis and a high‐resolution dynamically downscaled counterpart. The dynamical downscaled simulations, constrained using spectral nudging, enable the replication of historical events at a higher resolution. This allows the SRCNN to emulate dynamical downscaling effectively. Modifications, such as incorporating elevation data and data pre‐processing enhances overall model performance, while using exponential and quantile loss functions improve the simulation of extremes. Our findings show SRCNN models efficiently and skillfully downscale precipitation from GCMs. Future work will expand this methodology to downscale additional variables for future climate projections.

Labels

Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.